Skip to content
NBRHD TWINSPATIAL HOME PASSPORT
Launch demo 3D reconstruction examples

EPIC MEGAGRANTS PROJECT

NBRHD
Twin

Robot-assisted spatial passports for real homes.

We combine photorealistic spatial capture, structured 3D geometry and property data so a home can be explored, understood and reimagined — with Unreal Engine as the target interactive runtime.

Watch project overviewLaunch interactive prototype

Live web prototype · Early render proof · Robot-assisted capture in development

01 / CURRENT PROOF

From a live product prototype to a reconstructed real space

The project already combines a live property-discovery interface with real-space capture reconstructed as an early interactive 3D scene.

LIVE WEB PROTOTYPE

Neighborhood and property experience

A working web prototype for neighborhood-level property discovery, context and the move from map to immersive 3D exploration.

Launch prototype
REAL SPACE → STRUCTURED 3D

The same room, captured and reconstructed

Real footage on one side, an early 3D render on the other — entrance, furniture, windows and layout, matched to the source room.

Current Blender prototype validating room structure, navigation and interactions. Unreal Engine integration is the grant-funded next stage.

02 / THE IDEA

Photorealistic enough to experience.
Structured enough to edit.

  1. Photorealistic capture layer

    Gaussian splats preserve the real appearance, lighting, materials and visual detail of the property.

  2. Structured interactive layer

    Rooms, walls, openings, doors, furniture and surfaces become navigable, interactive and editable in Unreal Engine.

  3. Property data layer

    Measurements, orientation, sunlight, views, access, travel times and neighborhood context.

PHOTOREAL SPLATSTRUCTURED UNREAL SCENEPROPERTY DATASPATIAL HOME PASSPORT
We are building a hybrid twin: photorealistic splats for appearance, with a structured Unreal Engine layer for interaction, staging and renovation.

03 / ROBOT-TO-UNREAL PIPELINE

Robot-assisted capture

Our planned capture system uses a teleoperated quadruped carrying panoramic cameras, LiDAR and motion sensors through the property.

  1. 01

    Capture

    360° imagery · LiDAR · trajectory
  2. 02

    Reconstruct

    Camera alignment · geometry · semantic structure
  3. 03

    Build in Unreal Engine

    Navigation · interaction · materials · lighting
  4. 04

    Use

    Explore · understand · stage · renovate

PLANNED CAPTURE WORKFLOW

04 / PRODUCT EXPERIENCE

Explore, understand, reimagine

EXPLORE

Move from the neighborhood to the grounds and interior.

UNDERSTAND

See layout, orientation, views, sunlight and surrounding context.

REIMAGINE

Test furniture, materials, staging and renovation concepts.

05 / MEGAGRANT MILESTONE

What the MegaGrant unlocks

We already have access to a real Côte d’Azur villa for the pilot. The owner has agreed to make the home and grounds available for a controlled robot-assisted capture campaign.

Within six months, we will capture one real Côte d’Azur property and deliver a complete Unreal Engine vertical slice covering the home, grounds and selected neighborhood context. It will combine photorealistic capture, structured geometry and property data, with Original, Staged and Renovation states. Deliverables will include a packaged Unreal Engine build, a browser-accessible review experience and a documented anonymized workflow for the 3D community.

06 / FUTURE NETWORK

Toward a spatial property network

Long term, owner-controlled home passports can form a privacy-aware property network. Buyers will be able to search homes by real spatial characteristics, explore listed and off-market properties, and send qualified expressions of interest.

07 / TEAM

A team in physical systems and 3D vision

Vitalii Kapranov

FOUNDER & HEAD · CRAMBUZZ

Founder and head of Crambuzz, including aiprintgen.com. Background in aerospace robotics — hardware flown to the ISS — and drone systems, plus years of hands-on experience across real estate and machine learning that shape how NBRHD Twin turns spatial capture into usable property data.

Davit Piliposyan

SCIENTIFIC LEAD · DIGITAL MODELLING

Head of M:3L Lab. Research experience in applied mathematical modelling, mechanics and machine-learning-based engineering systems.

Danila Rukhovich

3D VISION & RECONSTRUCTION

Lead Researcher at M:3L and AI researcher at Google, specializing in 3D vision, point-cloud understanding, object detection and CAD reconstruction.

Physical laboratory transitioning into its digital reconstruction

NBRHD TWIN / EPIC MEGAGRANTS

From real homes
to persistent spatial assets.

Launch interactive demo Watch project overview